Ethical ends: Effect of abstract mindsets in ethical decisions for the greater social good

Organizational Behavior and Human Decision Processes 124 (2014) 110–121

Contents lists available at ScienceDirect

Organizational Behavior and Human Decision Processes

journal homepage: www.elsevier .com/ locate /obhdp

Ethical ends: Effect of abstract mindsets in ethical decisionsfor the greater social good

http://dx.doi.org/10.1016/j.obhdp.2014.02.0010749-5978/� 2014 Elsevier Inc. All rights reserved.

⇑ Corresponding author. Fax: +1 305 348 3792.E-mail addresses: [email protected] (J. Rixom), [email protected].

edu (H. Mishra).1 Fax: +1 801 581 3152.

Jessica Rixom a,⇑, Himanshu Mishra b,1

a Florida International University, RB-325, 11200 SW 8th Street, Miami, FL 33199, United Statesb David Eccles School of Business, University of Utah, SFEBB 7111, 1645 E. Campus Center Drive, Salt Lake City, UT 84112, United States

a r t i c l e i n f o a b s t r a c t

Article history:Received 29 April 2012Accepted 19 February 2014

Accepted by Madan Pillutla

Keywords:Ethical decisionsEthical principlesEthicsDishonestyHonestyDonationsConstrual level theoryDecision making

We explore the impact of construal level on decisions involving conflicts between multiple ethicalprinciples. Whereas abstract mindsets are associated with a focus on ethical issues and superordinateconcerns, concrete mindsets are associated with financial self-interest. With abstract mindsets, we findthat people abide by rather than violate ethical principles when only the self would benefit (single prin-ciple) but they violate ethical principles when doing so is a conduit for a greater social good (multipleprinciples). With concrete mindsets, people violate ethical principles for personal gain with less concernfor the impact on the greater social good. Specifically, with abstract mindsets, people were dishonest tosecure larger donations (Study 1) and dishonest to make larger (smaller) donations to charities that sup-ported (threatened) the greater social good (Study 2a, Study 2b) whereas with concrete mindsets, peoplefocused more on dishonesty for personal gain (Study 1, Study 2a, Study 2b).

� 2014 Elsevier Inc. All rights reserved.

Introduction

Ethical decision-making continues to capture the attention ofacademics and mainstream media alike. Regular reports, from theftand employee misconduct to elaborate Ponzi-schemes and Enron-like financial statement frauds, highlight the abundance ofdishonest behaviors that occur when people face ethical decisions.Oftentimes, these unethical behaviors are undertaken to increasepersonal gain, but at times, unethical behaviors occur for the sakeof the greater good.

There are many instances when unethical behaviors in one do-main are done to support a greater good in another domain, ratherthan for personal gain alone. In essence, these are situations whendoing something ‘‘wrong’’ enables people to do something ‘‘right.’’Consider, for example, burglars who stole 1000 documents from aPennsylvania FBI office in 1971. Peace activists wanting to reveal‘‘massive illegal surveillance and intimidation,’’ by J. Edgar Hooverbroke into the FBI office, stole files, and released them to the public(Isikoff, 2014). Here the burglary is clearly a dishonest act butexposing surveillance and intimidation is perceived as a greater

good. Stated differently, in this situation, dishonesty is a smallprice paid for the greater good. One can describe such exposé asa charitable act because the burglars did not reveal their identitiesfor more than 42 years; they did not gain any financial or fame-related benefits from the act of writing books. However, the exposédid curb some of the government’s antidemocratic practices. In an-other case of decisions where unethical behaviors directly benefit agreater good, computer hackers recently stole credit card numbersthat could have been used for selfish purposes, such as shoppingsprees, but instead were used to make large donations to variouscharities (Vinograd, 2011). In other words, one ethical principle(stealing is bad) was violated in order to directly support anotherethical principle (donating is good) that benefits a greater purpose.

When faced with situations in which violating an ethical princi-ple can support a greater good or oneself, one must decide onwhich ends to focus and whether the ends are justified by themeans. Rather than debating whether one should focus on thegreater good or on oneself, in the current research, we considerwhen individuals will focus on one aspect or the other. Consistentwith past work (e.g., Amit & Greene, 2012; Bartels, 2008; Broeders,van den Bos, Müller, & Ham, 2011; Toure-Tillery & Fishbach, 2012)suggesting that such decisions are not necessarily dispositionallyconstant but instead may be situationally triggered, we suggestthat abstract versus concrete mindsets influence whether one willfocus on benefits to a greater social good or benefits to oneself.

http://crossmark.crossref.org/dialog/?doi=10.1016/j.obhdp.2014.02.001&domain=pdf

http://dx.doi.org/10.1016/j.obhdp.2014.02.001

mailto:[email protected]



http://dx.doi.org/10.1016/j.obhdp.2014.02.001

http://www.sciencedirect.com/science/journal/07495978

http://www.elsevier.com/locate/obhdp

J. Rixom, H. Mishra / Organizational Behavior and Human Decision Processes 124 (2014) 110–121 111

Specifically, we assess the proposition that construal level the-ory can inform our understanding of when one will focus on out-comes that support a greater social good or on outcomes thatsupport oneself and the influence this has on honesty. To do so,we consider situations that involve a single ethical principle andsituations that involve a conflict between multiple ethical princi-ples in order to assess the impact of construal level on which prin-ciple will be prioritized. In other words, we assess the role ofabstract and concrete mindsets on whether people will focus onremaining honest, being dishonest to support a greater social good,or being dishonest to help oneself.

Research on construal level theory (Trope & Liberman, 2003)shows that people’s judgments and decisions differ based onwhether an abstract, high-level or concrete, low-level mindset isenacted. Past research demonstrates that an abstract mindset leadspeople to abide by ethical principles (Agerström & Björklund, 2009;Eyal, Liberman, & Trope, 2008). However, since an abstract mind-set also focuses one on the superordinate ends (Liberman & Trope,1998), we suggest that activating an abstract mindset will lead todifferent behaviors based on the impact on a greater social good.Specifically, we anticipate that by activating an abstract mindset,people will not behave dishonestly for personal gain alone, butwe do anticipate that they will be dishonest when doing so pro-vides more support for a greater social good whereas with a con-crete mindset, people will consistently emphasize thesubordinate focus of helping oneself. We suggest that this occursbecause with an abstract mindset, people focus on the superordi-nate ends or greater social good inferred from the situation andprioritize it over the principle of honesty.

This manuscript is organized as follows. First, we review litera-ture on ethical decisions where violating one ethical principlesimultaneously supports a different ethical principle that empha-sizes a greater social good. Then, we provide an overview of con-strual level theory and its ability to provide insights intosituations where one ethical decision is present and where thereis a conflict between two ethical principles. Next, we present threestudies based on the principles of honesty and supporting a greatersocial good. Finally, we discuss the theoretical and practical appli-cations of this research.

Theoretical overview

Ethical decision-making is a regular part of life. Oftentimes,these decisions involve a single ethical principle such as to stealor not to steal from one’s employer. At other times, two ethicalprinciples are intertwined such that a decision to adhere to or vio-late one ethical principle simultaneously affects a second ethicalprinciple such as to steal or not to steal in order to do more or lessto help those in need.2 When the situation involves a single ethicalprinciple, there is no conflict. However, when the situation involves aconflict between two ethical principles, then one principle will takepriority over the other.

While a conflict may exist between two ethical principles, thesituation may not constitute an ethical dilemma in the traditionalsense. Unlike the contexts addressed in the current research, moraldilemmas are defined as ‘‘situations in which no moral choice iswithout undesirable moral consequences’’ (Ditto & Liu, 2012, p.55). Classic examples include the trolley problem (Foot, 1967)

2 In the current research, we conceptualize unethical behaviors as those thatviolate hypernorms or widely held standards and principles of ethical behaviorsbased on social norms or laws (Donaldson & Dunfee, 1994; Umphress & Bingham,2011). Thus, a behavior that violates hypernorms is classified as unethical, even if theconsequences of the action support another ethical principle. We classify dishonesty,for instance, as unethical, regardless of whether it benefits or harms either the self oranother entity. That is, while the outcome may have ethical or beneficial overtones,the behavior itself is unethical if it violates hypernorms.

and footbridge problem (Thomson, 1985) in which one must de-cide whether to kill one person in order for several others to liveand the Heinz dilemma (Kohlberg, 1969) where a husband mustdecide whether to steal expensive medication in order to keephis wife alive. Moral dilemmas such as these pit deontologicaland consequentialist perspectives against each other, forcing peo-ple to decide whether one presumably unethical act (taking onelife or stealing) can be justified if it results in what most peoplewould consider greater ethical ends (saving multiple lives or curingone’s dying spouse). While traditional dilemmas such as these areimportant, we explore a different type of conflict.

The current research explores situations where ‘‘undesirablemoral consequences’’ (Ditto & Liu, 2012, p. 55) are not required.We lift this restriction. That is, while with traditional dilemmas,something negative must occur (e.g., one or many must die; steal-ing is required or a loved one will die), in the current research eth-ical principles do not have to be violated to help a greater socialgood but, violating an ethical principle does allow one to do evenmore for the greater social good. In other words, the conflict thatexists between two ethical principles is self-imposed rather thanbeing imposed solely by the situation. People are freed up to abideby both ethical principles but since there is a conflict (i.e., abide byboth principles or violate one principle to do even more good) theymust decide which one takes priority and that will ultimatelyinfluence behaviors.

Within such conflicts, the beneficiary may also differ such asviolating a principle for personal gain alone or for the sake of an-other. A white lie may be told not for one’s own gain but to helpanother person financially (Erat & Gneezy, 2012; Gneezy, 2005)or principles of fairness may be violated by unequally distributingresources in order to help a person to whom empathy is felt (Bat-son, Batson, et al., 1995; Batson, Klein, Highberger, & Shaw, 1995).Beyond an individual beneficiary, ethical principles may also beviolated for the sake of more abstract notions such as fairness. Re-cent work, for instance, has found that by manipulating wealth-based inequity experimentally (e.g., wealthy graders/poor problemsolvers), people will behave dishonestly in an attempt to do moreto restore equality (Gino & Pierce, 2009, 2010). The participants didnot have to violate any ethical principles. They could haveremained honest and done a little to restore equality (i.e., moder-ate earnings for both poor and wealthy solvers) but instead theyviolated the principle of honesty in order to do even more towardrestoring financial equality (i.e., dishonest to give more (less)money to poor (wealthy) problem solvers).

Pertinent to the current work, situational factors may alsoinfluence the extent to which one deems there to be value in pro-tecting a greater organizational good. In a recent conceptual paper,Umphress and Bingham (2011) posit that identification with anorganization and a desire to reciprocate an employers’ consideratetreatment are situational factors that influence whether employeesviolate ethical principles to do more to help an abstract, greaterorganizational good. Violating these principles can take manyforms from placing deceptive information into a bid in order to ob-tain a contract (Vardi & Wiener, 1996) to lying about one’s identityin order to gather competitive information. Flight attendants haveeven acknowledged lying to customers when they believed that itwould help protect the airline’s image (Scott, 2003). This deceitwas said to have taken place over concerns for the detriment thatcould occur to the greater organizational good if the company’sreputation was severely damaged (e.g., falling shareholder value,co-worker job loss), rather than concerns directly regarding per-sonal gain.

Across these scenarios, a common theme emerges; specifically,when there is only one ethical principle at play, people must decidewhether to abide by or violate the principle but when two ethicalprinciples present a conflict, people negotiate between potential

112 J. Rixom, H. Mishra / Organizational Behavior and Human Decision Processes 124 (2014) 110–121

outcomes and ultimately which principle takes priority over theother. While some people may not violate ethical principles to helpthemselves alone, they may desire to do even more for the sake of agreater social good and will do so even though it means violatinganother ethical principle. An emphasis on such ethical ends is afocus of both this type of decision and of an abstract mindset(Liberman & Trope, 1998). Therefore, we turn to construal leveltheory for insight into when people will adhere to ethical princi-ples and when they will violate these principles for either self-gainor for the benefit of a greater social good.

Construal level theory

Construal level theory argues that people mentally representevents in either abstract, high-level forms or in concrete, low-levelforms (Trope & Liberman, 2003). Focusing on abstract, global rep-resentations or overall gist of the situation, as opposed to concrete,specific representations, greatly influences judgments and deci-sions from categorization (Liberman, Sagristano, & Trope, 2002)and creativity (Förester, Friedman, & Liberman, 2004) to gambling(Sagristano, Trope, & Liberman, 2002) and politeness (Stephan,Liberman, & Trope, 2010) and has recently been extended intothe realm of moral decisions.

At first blush, it appears that construal level theory would pre-dict that with an abstract mindset, people would not violate anyethical principles, regardless of the outcome. For instance, past re-search that applied construal level theory to the context of ethicaland moral scenarios found that when a situation is abstract as op-posed to concrete in nature, ethical principles are salient and moreimportance is placed on abiding by these ethical principles (Agers-tröm & Björklund, 2009; Eyal & Liberman, 2012; Eyal et al., 2008).This occurs because moral principles represent the more generalelements of the situation and thus align more with an abstractrather than a concrete mindset (Eyal et al., 2008). Although thebasis of these findings come from judgments of another person’sactions rather than from the choices that one personally makeswhen faced with actual decisions involving ethical principles, itsuggests that activating an abstract mindset will lead people toabide by ethical decisions, regardless of the consequences.However, this prior research has only considered single ethicaldecisions such as whether another person should steal or not stealrather than situations that involve a conflict between two distinctethical principles, such as those explored in the current research.

When there is a conflict between adhering to ethical principlesin order to do a little for a greater social good or violating ethicalprinciples in order to do a lot for a greater social good, the greatersocial good benefits regardless, albeit to different degrees. To illus-trate, if honesty leads to a donation to an admirable charity but dis-honesty leads to an even larger donation, then adhering to theprinciple of honesty would still result in a donation but violatingthe principle of honesty would lead to an even larger donation.Past work suggests that when faced with one ethical decisionwhere there is no conflict, an abstract mindset will lead one to ad-here to ethical principles and to remain honest (Agerström &Björklund, 2009; Eyal, Liberman, & Trope, 2008) but it is silent onwhat approach will be used when there is a conflict between eth-ical principles. We suggest that when there are competing ethicalprinciples, one principle will take priority over the other; specifi-cally, the principle associated with helping the greater social goodwill take precedence over or be at a higher level than the principleof honesty.

Emphasizing the ends or superordinate goals as opposed to themeans or subordinate aspects of a situation are central compo-nents of abstract mindsets (Fujita, Henderson, Eng, Trope, &Liberman, 2006; Liberman & Trope, 1998; Trope & Liberman,2003). The ‘‘why’’ of the situation is more relevant when an ab-

stract rather than a concrete mindset is activated. Thus, whenthe outcome of a decision has the potential to benefit a greater so-cial good, these ends are likely to be more salient with an abstractrather than a concrete mindset. In alignment, taking a distant, ab-stract perspective activates the idealistic self, which emphasizeschoices that conform to ethical principles and values rather thanto pragmatic concerns (Kivetz & Tyler, 2007). When more thanone ethical principle is involved, the principles that do the mostto support the idealistic self may take precedence.

It has been suggested that when ethical principles are viewed as ahierarchy, the notion of helping a greater or bigger purpose may se-cure a higher, more abstract place in the hierarchy relative to otherprinciples (Broeders et al., 2011). For instance, framing moral dilem-mas in the context of ‘‘saves lives’’ was found to align more with anabstract mindset so that when the event is said to occur in the dis-tant, abstract future, people report a greater likelihood of lettingone person die in order to save many lives (Broeders et al., 2011).Accordingly, the opportunity to provide more support to a greatersocial good may align more closely with an abstract mindset andprovide the justification necessary to violate another ethical princi-ple. That is, the greater social good serves as a superordinate, focalconcern that overshadows the steps required to achieve those ends.

We anticipate that when an abstract mindset is activated, peo-ple will not violate ethical principles for personal gain alone butwill violate ethical principles when doing so enables them to doeven more for a greater social good. When only the self will benefit,there is just one ethical principle at stake, for instance, to lie or notto lie. There is no conflict and thus no hierarchy of ethical deci-sions. The ‘‘why’’ of the situation—to help oneself—may appearinsufficient to justify violating the only ethical decision at stake,so activating an abstract mindset will lead people to abide by eth-ical principles. This aligns with the predictions of earlier workinvolving judgments of ethical scenarios through the lens ofconstrual level theory.

On the other hand, when there is a conflict such that violatingan ethical principle allows one to do more for a greater social good,people must decide whether to adhere to or violate that principle.With an abstract mindset, people focus on the superordinate‘‘why’’ of the situation (Fujita, Henderson, et al., 2006; Liberman& Trope, 1998) which may overshadow other considerations. Inthe context of ethical decisions, this superordinate element—todo even more for a greater social good—may overshadow or justifythe decision to violate an ethical principle.

That is, while both charity and honesty are ethical principles,charity involves going one-step further by requiring one to makesacrifices for others. The more cherished the object one gives up,the more selfless the act becomes. In an abstract mindset, whenhonesty is pitted against charity, honesty becomes the sacrificethat one can make to help others. Here one sacrifices one’s ownvalued ethical principle of honesty to help a greater social good.In sum, following only an honest approach in an abstract mindsetmay not appear as selfless or justifiable as being dishonest to help agreater social good. Accordingly, we anticipate that when ethicalprinciples conflict in this manner, activating an abstract mindsetwill lead to violating ethical principles specifically for the benefitof the greater social good with less regard for whether the decisionsimultaneously harms, helps, or has no impact on one’s ownpersonal outcome because the emphasis is focused outwardly onbenefiting a greater good rather than inwardly on oneself.

To this point, our focus has been on the anticipated change inbehaviors when an abstract mindset is activated, but what abouta concrete mindset? When a concrete mindset is activated, wedo not expect much fluctuation in behaviors as a function of theimpact on a greater social good. Instead, we anticipate that anabstract mindset and a concrete mindset will lead to different pro-cesses. In other words, while we anticipate that the beneficiary


may moderate decisions about violating ethical principles when anabstract mindset is activated, we do not anticipate much variationwhen a concrete mindset is activated.

With a concrete mindset, less emphasis is placed on ethicalprinciples (Agerström & Björklund, 2009; Eyal & Liberman, 2012;Eyal et al., 2008) so the greater social good is unlikely to hold asmuch weight as it would with an abstract mindset. This limited fo-cus on ethical implications is supported by findings that peopleusing a near, concrete view are more likely to construe events interms that are ‘‘devoid of moral implications’’ (Eyal & Liberman,2012, p. 189). Instead, the emphasis is placed on subordinate orproximate concerns (Fujita, Henderson, et al., 2006; Liberman &Trope, 1998). So while notions of supporting a greater good aremore closely aligned with an abstract mindset (Eyal et al., 2008),financial self-interest is more aligned with a concrete mindset(Hunt, Kim, Borgida, & Chaiken, 2010; Kivetz & Tyler, 2007) andthus may stand out more as a concern of interest. With the impactof one’s actions on a greater social good registering less importancewith a concrete relative to an abstract mindset, behaviors in thedecision contexts that we explore are more likely to remain stea-dily in favor of self-interest when a concrete mindset is activated.While most people would like to view themselves as honest(Toure-Tillery & Fishbach, 2012), it is found that people generallyengage in some level of dishonesty when given the opportunity(Mazar, Amir, & Ariely, 2008; Shalvi, Dana, Handgraaf, & De Dreu,2011). For instance, people believe the actions taken at the begin-ning and end of a sequence are more indicative of personal stan-dards and so they are more likely to adhere to ethical principlesat the beginning and end of a sequence but to violate those sameethical principles in the middle of the sequence (Toure-Tillery &Fishbach, 2012). This allows them to walk away from the situationfeeling positive about their ethical self-image while still experienc-ing the personal-benefits that came from violating ethical princi-ples (e.g., cheating to save time). While the current research doesnot involve decision sequences, with a concrete mindset, peopleare also more lenient toward violators of ethical principles (Eyalet al., 2008), so we anticipate that dishonesty will occur more reg-ularly when a concrete mindset is activated. We further suggestthat since self-interest is a more subordinate and proximate con-cern than helping a greater social good, when a concrete mindsetis activated, people will engage in behaviors with consequencesthat are most likely to benefit the self.

To summarize, we anticipate that activating an abstract relativeto a concrete mindset will lead to different decisions when facedwith single or multiple ethical principles. We propose that whenpeople face a decision involving a single ethical principle and anabstract mindset is activated, they will not violate the ethical prin-ciple for personal gain alone. However, we do not anticipate that anabstract mindset will always lead to a strict adherence to ethicalprinciples. Instead, when there is a conflict between two ethicalprinciples, activating an abstract mindset will lead people to prior-itize one principle over the other. More specifically, they will focuson the greater social ends inferred from the situation and will pri-oritize doing even more to help the greater social good over theprinciple of honesty. Violating the ethical principle of honesty willoccur in the direction that provides the most support for the great-er social good, whether this harms, helps, or has no impact on theperson’s own outcomes. When a concrete mindset is activated,however, we anticipate that people will violate the ethical princi-ple of honesty to roughly the same extent in the direction that pro-vides the most personal gain. In essence, with a concrete mindset,people will emphasize outcomes that concern personal gain ratherthan focusing heavily on the impact on a greater social good. Weassess these propositions across three studies.

Current studies

The intent of our studies is to assess the notion that an abstractrelative to a concrete mindset differentially influences behaviorsfor ethical decisions that affect a greater social good or oneself.In other words, we evaluate the role of construal level theory inaccounting for why people who have an opportunity to support agreater social good without violating any ethical principles mayinstead decide to violate an ethical principle in order to do evenmore for the same social good. We suggest that a focus on thesuperordinate or ultimate ends associated with the greater socialgood—a feature of an abstract mindset—will take priority overthe principle of honesty. Therefore, across studies, we activateeither an abstract or a concrete mindset using established primes(Freitas, Gollwitzer, & Trope, 2004; Fujita, Trope, Liberman, &Levin-Sagi, 2006) and we assess actual behaviors involving realmoney in Study 1 and Study 2a.

While we emphasize situations involving a conflict between theprinciples of honesty and doing more to help a greater social good,we also include conditions in which dishonesty only benefits one-self. These conditions present a single ethical decision (to be hon-est or dishonest) when only the self will benefit from dishonesty.This serves two purposes. First, it allows us to see whether actualbehaviors are consistent with the patterns associated with past re-search. Second, it enables us to assess whether introducing a great-er social good into the mix does lead to behavior changes based onthe beneficiary in the context of construal level.

We use an established cheating paradigm (Mazar et al., 2008) toassess whether people abide by the principle of honesty when dis-honesty affects one’s own financial gain, a greater social goodthrough a monetary donation, or both. Across the studies, the taskdesign allows the participants to be honest and still support thegreater social good such as supporting an admirable charity orwithholding funds from an unadmirable charity. However, partici-pants also have the opportunity to be dishonest, which can lead toa larger benefit for the greater good either through donating evenmore to the admired charity (Study 1 and Study 2a) or even less tothe unadmired charity (Study 2a).

Payments to the participants or charities were based on actualor stated performance (depending on condition) on a timed task.Providing actual monetary incentives increased the relevance ofthe ethical decisions in a non-hypothetical context. The Make-A-Wish Foundation received all monetary donations made through-out the studies.

Study 1

The primary aim of Study 1 is to assess the impact of activatingeither an abstract or a concrete mindset on ethical decision-mak-ing when the beneficiary is oneself relative to an admired charityserving a greater good. We assess this in the context of honestyusing a cheating paradigm (Mazar et al., 2008). In all conditions,the monetary payment to participants was dependent on their ac-tual or stated performance on a timed puzzle task. Dishonesty wasinferred if the stated performance of a condition in which dishon-esty was possible was significantly higher or lower than the actualperformance of the control conditions in which dishonesty was notpossible (Mazar et al., 2008; Mead, Baumeister, Gino, Schweitzer, &Ariely, 2009).

Using this context, participants decide whether to be honest ordishonest when the decision influences their own earnings orwhen the decision influences the amount donated to theMake-A-Wish Foundation, an admirable charity that supports agreater social good (helping terminally ill children), dependingon condition. We anticipate that when faced with the single ethical

3 In this priming task, a series of 40 words were presented two at a time toparticipants on the computer. In the abstract prime condition, participants generateda category label for each provided word. They were instructed to ask themselves thequestion, ‘‘[Provided word] is an example of what?’’ and to come up with the mostgeneral category to which each word belonged. For instance, for the provided word,‘‘dog,’’ participants might type, ‘‘animal’’ as the category label. Participants in theconcrete prime condition came up with a word that was an example of the providedword. This was done for the same series of 40 words by answering the question, ‘‘Anexample of [provided word] is what?’’ They were instructed to generate the mostspecific example possible. For instance, for the provided word, ‘‘dog,’’ participantsmight type, ‘‘poodle’’ as the example.

4 When the provided word was an example of the participant’s response (e.g.,participant indicated ‘‘animal’’ when the provided word was ‘‘dog’’), the response wascoded as 1. When the participant’s response was an example of the provided word(e.g., participant indicated ‘‘poodle’’ when the provided word was ‘‘dog’’), theresponse was coded as �1. Responses not fitting either criterion were coded 0.


principle of whether to be honest or dishonest for personal gainalone, participants primed with an abstract mindset will abide bythe principle of honesty. They will not be dishonest for personalgain alone. However, when there is a conflict between the principleof honesty and the desire to do even more to support a greater so-cial good, we propose that when an abstract mindset is activated,the greater social good will be prioritized over the desire to remainhonest. In other words, while we anticipate that when an abstractmindset is primed, participants will not be dishonest for personalgain alone, we do anticipate that they will be dishonest whendoing so enables them to provide additional support for a greatersocial good, in this case, the Make-A-Wish Foundation. When aconcrete mindset is primed, however, we do not anticipate aswitch in behaviors based on the beneficiary and instead anticipatethat concerns over personal gain will be prioritized, leading toroughly the same level of dishonesty across beneficiaries.

Method

Participants and design

Two hundred ninety-one undergraduate students (mean age23.92 years, 36.9% female) participated in the study for partialcourse credit. In Study 1, participants learned at the beginning ofthe study that they would earn money either for themselves orfor a charity, depending on condition. Therefore, it was feasiblethat in the control conditions in which dishonesty was not possi-ble, actual performance could still vary based on the beneficiary(i.e., extend more/less effort for the charity relative to the self).While we did not predict a priori (nor did we find) any differencesin actual performance among the control conditions, we accountedfor this possibility by including control conditions that earnedeither for themselves or for the admirable charity. Accordingly,participants were randomly assigned to one of eight between-participants conditions within a 2 (CLT prime: abstract vs. con-crete) � 2 (beneficiary: self vs. charity) � 2 (dishonesty: not possi-ble (control) vs. possible) design. Earnings were actual money paiddirectly to the participant (self-benefit conditions) or to the charity(charity-benefit conditions). Participants in the control conditionswere unable to dishonestly over- or under-report performance.The dependent variable was the actual (control conditions) or sta-ted (dishonesty possible conditions) number of correctly solvedpuzzles on a timed task. Dishonesty was inferred when the averageperformance in a condition in which dishonesty was possible wassignificantly higher or lower than the control.

Procedure

Upon arrival, participants sat at desks with a computer and readthat they would complete a series of unrelated paper- and com-puter-based tasks. First, participants read that they would com-plete a timed puzzle task. Each solved puzzle was worth 25cents, up to $5 total. In the self-benefit conditions, participantsread that they would earn money for themselves. In the charity-benefit conditions, participants read that they would earn moneyfor the Make-A-Wish Foundation.

Next, participants took out a clipboard with a sheet of paperthat had 20 matrices, each with 12 three-digit numbers (Mazaret al., 2008). They had 5 min to find the two numbers on each ma-trix that added up to 10. When the 5 min were over, participants inthe conditions where dishonesty was possible saw the correct an-swers on the computer and graded their own performance. Theparticipants in the control conditions did not grade their own per-formance and did not have an opportunity to be dishonest. All par-ticipants then read that payment would occur at the end of thestudy and were instructed to set the clipboard with the matrix

sheet aside. Then, they proceeded to an ostensibly different taskin which they completed an abstract or concrete mindset prime.

To activate either an abstract, high-level or a concrete, low-levelmindset, participants completed a construal level prime (Fujita,Trope, et al., 2006).3 After the mindset prime, participants read pay-ment instructions, which varied by condition. The instructions in thecontrol conditions asked participants to give the matrix sheet to theresearcher. Dishonest behavior was not possible because the re-searcher directly graded the matrix sheet. Earnings were handed tothe participant (self-benefit) or earnings were placed in an envelope(charity-benefit) for the Make-A-Wish Foundation in front of theparticipants.

Dishonesty was possible in the other conditions because partic-ipants graded themselves, destroyed the evidence, and paymentwas for the number of matrices that they reported solving. Partic-ipants in the self-benefit (charity-benefit) conditions read a remin-der that they (the Make-A-Wish Foundation) would earn a quarterfor each correctly solved matrix for a maximum of $5. After readingthe payment instructions, participants went behind a partition, outof the researcher’s view. While behind the partition, they countedthe number of correctly solved matrices, tore up the matrix sheet,threw the pieces into a recycling bin, and indicated the number ofsolved puzzles and the payment amount on a form. Participantsthen handed the payment form to the researcher. In the self-bene-fit conditions, the researcher handed the amount indicated on thepayment form to the participant. In the charity-benefit conditions,the researcher assured the participants of the donation and placedthe earnings in an envelope. The donation to the Make-A-WishFoundation occurred after the data collection period.

Finally, participants returned to the computer to answer demo-graphic questions. In the charity-benefit conditions, participantsalso indicated the extent to which they believed that the donationwould be made to the Make-A-Wish Foundation on a scale from 1(No, definitely will not be made) to 7 (Yes, definitely will be made).

Results and discussion

Of the original 291 participants, four participants were ex-cluded from the analyses due to technical problems. Data fromthe remaining 287 participants are included in the followinganalyses.

Manipulation checks

Two independent judges evaluated the level of construal foreach participant based on their responses in the priming task(Fujita, Trope, et al., 2006).4 Disagreements were resolved by discus-sion (initial agreement was 95.5%). Higher total scores indicate ahigher-level, more abstract mindset. The analysis revealed that theprime successfully manipulated construal level. Participants whogenerated category labels had significantly more abstract responses


(M = 31.12) than those who generated exemplars (M = �30.36),t(285) = 34.88, p < .001.

Among participants earning for the charity, there were no dif-ferences between conditions in terms of the belief that the moneywould be donated to the Make-A-Wish Foundation,F(3,141) = .071, p = .975, and the average believability was signifi-cantly above the scale mid-point (M = 6.09), t(144) = 18.25,p < .001. To maintain anonymity, participants’ payment sheetsand their answers on the computer were not associated with oneanother so we are unable to assess the number of correct answers(actual or stated) as a function of donation believability.

Number of matrices solved (honesty)

To assess whether there were any unintended differencesamong the control conditions, we first analyzed the total numberof solved matrices using a 2 (CLT prime: abstract vs. concrete) � 2(beneficiary: self vs. charity) � 2 (dishonesty: not possible (con-trol) vs. possible) ANOVA. The analysis revealed a significantthree-way interaction F(1,279) = 3.711, p = .055, g2

p = .013. We didnot anticipate differences within the control conditions and thelack of a significant interaction between dishonesty and benefi-ciary (p = .637) and between dishonesty and CLT prime (p = .411)suggested that this was the case. To examine directly whetherthere were differences in actual performance between the fourcontrol conditions, we conducted a 2 (CLT prime: abstract vs. con-crete) � 2 (beneficiary: self vs. charity) ANOVA. As anticipated,within the control conditions, there were no significant differencesbased on CLT prime or beneficiary. Since our primary interest is theimpact of construal level on decisions involving ethical principlesand there are no significant differences within the control condi-tions, we collapse across beneficiary, combining those who com-pleted the abstract prime into one control group and those whocompleted the concrete prime into another control group.

Next, we turned to the primary analysis to assess the impact ofconstrual level on honesty when the self or the charity would ben-efit by conducting a 2 (CLT prime: abstract vs. concrete) � 3(deception: not possible (control) vs. possible for self-benefit vs.possible for charity-benefit) ANOVA. The analysis revealed a signif-icant interaction, F(2,281) = 3.12, p = .046, g2

p = .022 and a signifi-cant main effect of deception condition, F(2,281) = 6.67, p = .001,

4.08a 4.15a

6.00

4.315.25

5.92

0

1

2

3

4

5

6

7

8

Concrete Abstract

Num

ber o

f M

atric

es

Error Bars: +- 2 SE

Control Self-Benefit Charity-Benefit

Fig. 1. Influence of construal level on number of matrices solved (control) orreported to have been solved correctly in Study 1 as a function of construal levelmindset (abstract or concrete) and payment condition (control, self-benefit,charity-benefit). aControl conditions were collapsed across beneficiary. There wereno significant differences in the concrete mindset condition between the number ofmatrices actually solved when the self (M = 3.94) or the charity (M = 4.22) benefitedor in the abstract mindset condition when the self (M = 4.19) or the charity(M = 4.11) benefited.

g2p = .045 (see Fig. 1). Planned contrasts provided additional insight.

First, consistent with our previous analysis, when dishonesty wasnot possible (control), there were no significant differences in thenumber of correctly solved matrices based on construal level(MAbsControl = 4.15 vs. MConControl = 4.08), F(1,281) = .04, p = .834.

Next, we assessed our earlier predictions when dishonesty waspossible. Dishonesty was inferred when the average claimed per-formance of a given condition was significantly higher or lowerthan the respective control. In alignment with the predictions ofprevious literature, we found that with an abstract mindset, therewere no significant differences in the number of matrices reportedsolved between the control condition (M = 4.15) and the conditionin which dishonesty was possible for personal gain (M = 4.31),F(1,281) = .12, p = .734, suggesting that participants were honestin their reporting. However, with a concrete mindset, the averagenumber of matrices reported solved was significantly higher(M = 6.00) compared to the control condition (M = 4.08),F(1,281) = 6.12, p = .018, g2

p = .056 suggesting that participantsoverreported performance. Consistent with our expectations, whenit was possible to be dishonest in order to influence the amount do-nated to an admirable charity, participants primed with an abstractmindset reported solving significantly more matrices (M = 5.92)than those in the control condition (M = 4.15), F(1,281) = 5.76,p = .021, g2

p = .051. When a concrete mindset was activated, thenumber of matrices reported solved did not significantly differbased on the beneficiary (MConSelf = 6.00 vs. MConCharity = 5.25),F(1,281) = .58, p = .449.

Finally, we explored the proportion of participants within eachcondition who reported solving (actually solved for control condi-tions) a low number (0–2), average number (3–5), high number (6–10), or very high number (11–20) of matrices (see Table 1). Wefound that when an abstract mindset was activated, only 25.0%of participants reported solving six or more matrices when dishon-esty would lead to personal gain but that 41.6% reported solvingsix or more matrices when dishonesty would provide a largerdonation to the Make-A-Wish Foundation, with most of this highreporting falling into the six to ten matrices range. These findingssupport prior research suggesting that when people are dishonest,they tend to be somewhat rather than completely dishonest(Mazar, Amir, & Ariely, 2008; Shalvi, Dana, Handgraaf, & De Dreu,2011).

Together, these findings are consistent with the notion thatwith an abstract mindset, a focus on supporting a greater socialgood is prioritized over the desire to remain honest. Specifically,we find that when an abstract mindset is primed, people are dis-honest to provide even more money to a charity that supports agreater social good, in this case, helping terminally ill children,but not to help themselves alone. Based on the control condition,honest reporting by those primed with an abstract mindset wouldlead to an average donation of $1.04 but when there was a conflictbetween remaining honest and doing even more for the greatersocial good, the greater good was prioritized and dishonest over-reporting led the average donation to increase to $1.48. Consistentwith propositions from prior research, we also find that whendishonesty would lead to personal financial gain, activating anabstract (concrete) mindset resulted in honesty (dishonesty).Additionally, when activating a concrete mindset, we did not seesignificant changes in the level of dishonesty based on the benefi-ciary, which aligns with a connection between a concrete mindsetand self-interest rather than social concerns.

Study 2

In Study 2a, we replicate and extend the findings from Study 1and in Study 2b, we explore the ends on which people with anabstract or a concrete mindset focus.

Table 1Influence of construal level on the proportion of participants who solved (control) or reported to have solved correctly a certain number of matrices in Study 1 as a function ofconstrual level mindset (abstract or concrete) and payment condition (control, self-benefit, charity-benefit).

Conditions

Concrete Control Abstract Control

Number of solved matrices Self-benefit(n = 35)

Charity-benefit(n = 36)

ConcreteControlcollapseda

(n = 71)

Self-benefit (n = 36) Charity-benefit(n = 37)

Abstractcontrolcollapseda

(n = 73)

Concreteself-benefit(n = 35)

Abstractself-benefit(n = 36)

Concretecharity-benefit(n = 36)

Abstractcharity-benefit(n = 36)

0–2 Matrices 31.4% 13.9% 22.5% 11.1% 13.5% 12.3% 22.8% 25.0% 30.6% 16.7%3–5 Matrices 42.9% 61.1% 52.1% 72.2% 67.6% 69.9% 28.6% 50.0% 22.2% 41.7%6–10 Matrices 25.7% 25.0% 25.4% 16.7% 18.9% 17.8% 40.0% 25.0% 38.9% 33.3%11–20 Matrices 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 8.6% 0.0% 8.3% 8.3%

Mean matrices solved 3.94 4.22 4.08 4.19 4.11 4.15 6.00 4.31 5.25 5.92

a In the control conditions, dishonesty was not possible. Control conditions were collapsed across beneficiary.


Study 2a

Study 2a has two primary objectives. The first objective is toinvestigate decisions when payments to both the participant andcharity are affected concurrently by honest or dishonest behaviors.This enables us to assess our proposition that activating an abstractmindset will lead to dishonest behaviors for the sake of the greatersocial good and that this will hold even when doing so simulta-neously leads to lower or higher personal gain. This design also al-lows us to assess the proposition that with a concrete mindset, theemphasis is placed on personal gain, leading to dishonesty primar-ily to financially benefit oneself.

In Study 1, participants in the charity-benefit conditionsworked directly on behalf of the charity. All of their actual or statedearnings went to the charity so the decision to be honest or dishon-est had no impact on the participant’s personal finances. In Study2a, we introduce a trade-off such that participants receive moneyfor each solved puzzle and a charity receives money for eachunsolved puzzle. Thus, reporting a larger (smaller) number ofsolved puzzles results in more (less) money for the participantand simultaneously less (more) money being donated to the char-ity. Participants who dishonestly underreport performance willearn less while simultaneously making a larger donation and thosewho dishonestly overreport performance will earn more whilesimultaneously making a smaller donation.

The second objective is to evaluate the impact of construal levelon honesty when dishonesty affects the donation amount for acharity that serves an admirable cause (good charity) or acontemptible cause (bad charity). The ‘‘good’’ charity was theMake-A-Wish Foundation and the ‘‘bad’’ charity was the fictitiousResource Use Fund, which was said to pay legal expenses for thosecharged with environmental damage. Securing larger donations foran admirable charity would serve the greater social good. Con-versely, in the bad charity condition, ensuring smaller donationsfor the charity would serve the greater social good. In both cases,dishonesty—albeit in different directions—further facilitates a great-er social good.

The design of Study 2a enables us to assess the proposed switchwhen an abstract mindset is activated from adhering to the princi-ple of honesty when it is the only principle of concern to a focus ona greater social good when there is a conflict between helping thegreater good and remaining honest. Strict adherence to ethicalprinciples would lead to honesty, regardless of the impact on agreater social good. When, however, a greater social good is prior-itized over honesty, we would find a pattern of underreporting per-formance in order to provide even more support for the good,admirable charity and overreporting performance in order to give

less support to the bad, unadmirable charity that threatens thegreater social good.

Consistent with the findings from Study 1, we anticipate thatwhen only the self will benefit from dishonesty, participantsprimed with an abstract mindset will remain honest, that is, theaverage number of correctly solved puzzles will not significantlydiffer from that of the control. When there is a conflict betweenhonesty and providing more support for a greater social good,however, we anticipate that when an abstract mindset is activatedthe greater social good will be given priority over honesty. In otherwords, when an abstract mindset is activated, we anticipate thatparticipants will underreport (overreport) performance, leadingto more money being given to (withheld from) the good (bad)charity which simultaneously leads to smaller (larger) personalearnings relative to the control. When a concrete mindset is acti-vated, we anticipate that the emphasis will be placed on personalgain and that on average participants will overreport performancefor personal gain whenever dishonesty is possible, whether the dis-honesty affects only oneself, a good charity, or a bad charity.

Method


Two hundred seventy undergraduate students (mean age23.91 years, 33.7% female) took part in the study for partial coursecredit. Participants were randomly assigned to either an abstract ora concrete prime condition and to one of four payment conditions:control, self-benefit only, self-benefit with remaining donated tothe good charity, or self-benefit with remaining donated to thebad charity.

Study 2a used the same matrix task as in Study 1 but the pay-ment instructions, timing of information about the charity, andconstrual level prime differed. As in the previous study, the depen-dent variable was the actual (control conditions) or reported (allother conditions) number of correct answers and dishonesty wasinferred when the average performance of a given condition wassignificantly higher or lower than the control.

Procedure

Participants sat at individual desks and read on the computerthat they would be completing a few unrelated paper- and com-puter-based tasks. The instructions informed participants that theywould complete a timed puzzle task. They would earn 25 cents foreach solved puzzle, up to $5 total. In each condition, the partici-pants received these earnings.


Each participant then had 5 min to complete the same matrixtask described in the Study 1 procedure section. Following the ma-trix task, participants in the conditions in which dishonesty waspossible—but not in the control—saw the correct answers on thecomputer and graded their own performance. All participants readthat payment would take place at the end of the study and wereasked to set aside the matrix sheet. Next, participants started aseemingly different task in which they completed an establishedabstract or concrete mindset prime.

The abstract, high-level or concrete, low-level construal primebased on Freitas et al. (2004) was completed on the computer.5 Fol-lowing the priming manipulation, participants read the paymentinstructions, which varied by condition. In the control condition, dis-honesty was not possible because participants handed the matrixsheet to the researcher who then graded the matrices and paid theparticipants based on their actual performance.

In the other conditions, dishonesty was possible. Participantsgraded themselves, destroyed the evidence, and were paid forthe number of matrices they reported solving. The details of thepayment, however, differed from those used in the previous study.In the self-benefit-only conditions, there was a reminder tellingparticipants that they would earn a quarter for each correct answerfor a maximum of $5. No mention was made of what wouldhappen to the unearned money. Participants then went behind apartition outside of the researcher’s view, counted the number ofcorrectly solved puzzles, tore up the matrix sheet, threw the piecesinto the recycling bin, and wrote both the number correct and theamount to be paid on a payment form. They then handed the pay-ment form to the researcher who paid them the amount on theform.

In the donation conditions, participants read that $5 had beenset aside. They would still earn a quarter for each correct answerbut the unearned portion of the $5 would be donated to charity.This was the first time that the charity was mentioned to partici-pants. While not explicitly stated, claiming few (many) correctlysolved puzzles would result in smaller (larger) personal earningsand simultaneously in larger (smaller) donations. The charitieswere the Make-A-Wish Foundation (good charity) or the ResourceUse Fund (bad charity). Following the instructions, participantswent behind the partition, counted the number of solved puzzles,destroyed and disposed of the matrix sheet, and filled out the pay-ment form. On the payment form, they indicated the number ofsolved puzzles, the amount they were to be paid, and the amountto be donated, which summed up to $5. Participants then gave thepayment form to the researcher who paid them the indicatedamount and placed the donation in an envelope labeled with thecharity’s name. All donations were given to the Make-A-WishFoundation after the data collection period.

Participants then went back to the computer to answer demo-graphic questions. The participants in the charity conditions alsoindicated the extent to which they believed that the donationwould be made to the Make-A-Wish Foundation or the ResourceUse Fund, depending on condition, using a scale from 1 (No,

5 To activate an abstract mindset, participants read a brief passage about ‘‘why’’somebody would participate in a research study. Then they saw the question, ‘‘Whydo I maintain good physical health?’’ on the top of the next screen. At the bottom ofthe screen was the statement, ‘‘Improve and maintain physical health.’’ Connectingthe question at the top and the statement at the bottom were four empty boxes withupward-pointing arrows labeled, ‘‘Why?’’ Participants filled in the bottom box andworked up to the top box. In the concrete mindset conditions, participants brieflyread about ‘‘how’’ somebody would participate in a research study. On the next screenthey saw the statement, ‘‘Improve and maintain physical health,’’ at the top and thequestion, ‘‘How do I maintain good physical health?’’ at the bottom. Connecting thestatement at the top and the question at the bottom were four empty boxes withdownward-pointing arrows labeled, ‘‘How?’’ Participants filled in the top box andworked down to the bottom box.

definitely will not be made) to 7 (Yes, definitely will be made). Par-ticipants in the bad charity condition also indicated, to the best oftheir knowledge, whether the Resource Use Fund was a real non-profit organization (No or Yes). Following the data collection peri-od, all participants were debriefed and assured that all donationswere provided to the Make-A-Wish Foundation.


Manipulation checks

Two independent judges coded the ‘‘why’’ and ‘‘how’’ responses(Fujita, Trope, et al., 2006).6 Disagreements were resolved by discus-sion (initial agreement was 96%). The analysis revealed that theprime successfully manipulated construal level. Participants who re-sponded to ‘‘why’’ prompts generated statements that reflected anabstract mindset (M = 3.63) compared to those who responded tothe ‘‘how’’ prompts (M = �3.53), t(268) = �79.89, p < .001.

Belief in the donation was high and significantly above the scalemid-point in both the good, Make-A-Wish Foundation (M = 6.47),t(72) = 18.83, p < .001, and the bad, Resource Use Fund (M = 6.15),t(59) = 11.79, p < .001 conditions. The extent to which participantsbelieved that the donation would be made did not significantly dif-fer based on whether the charity involved was good (M = 6.47) orbad (M = 6.15), t(131) = 1.44, p = .153. Results also show that91.7% of the participants in the bad charity condition indicated thatto the best of their knowledge, the Resource Use Fund was a realnonprofit organization. This did not significantly differ for the ab-stract (M = 90.0%) relative to the concrete (M = 93.3%),v2(1) = .218, p = .640, mindset conditions.7 Together, these findingslend support that the charity manipulations were successful acrossconditions.

Number of matrices solved (honesty)

An ANOVA revealed a significant interaction, F(3,262) = 12.72,p < .001, g2

p = .127, and significant main effects of both construal le-vel, F(1,262) = 8.01, p = .005, g2

p = .03, and payment condition,F(3,262) = 35.16, p < .001, g2

p = .29, (see Fig. 2). We then assessedour planned contrasts to gain further insight based on our expecta-tions. When dishonesty was not possible (control), the number ofcorrectly solved puzzles did not vary based on construal level(Mconcrete = 4.06 vs. Mabstract = 4.03), F(1,262) = .003, p = .953. Next,we assessed the extent and direction of dishonest behavior whendeception was possible. Dishonesty was inferred when the averagestated performance of a given condition was significantly lower(underreporting) or higher (overreporting) than the respectivecontrol. When faced with the decision of whether to be honest ordishonest for personal gain alone, when an abstract mindset wasactivated, there was no significant difference between the numberof matrices reported solved (M = 4.12) and the control condition(M = 4.03), F(1,262) = .03, p = .87, suggesting that participants hon-estly reported performance in the self-benefit-only condition.However, when a concrete mindset was primed, the number ofcorrectly solved matrices that was reported (M = 6.42) was signif-icantly higher than the number solved in the control condition(M = 4.06), F(1,262) = 11.36, p = .001, g2

p = .147, suggesting thatoverreporting occurred in the self-benefit-only condition. Consis-tent with Study 1 and previous research, participants with an

6 Responses superordinate to the preceding ones were coded 1 and subordinateresponses were coded �1. Responses not fitting either criterion were coded 0.

7 To maintain anonymity, there were no identification codes linking participants’payment forms to their computer-based answers; therefore, we can assess thesefollow-up questions on an aggregate level by condition, but we cannot explore howthese answers correspond to actual behaviors on an individual level.

4.06 4.036.42

4.12

6.95

2.36

8.4710.73

0

2

4

6

8

10

12

14

Concrete Abstract

Num

ber o

f M

atric

es

Error Bars: +- 2 SE

Control Self Only Self and Good Charity Self and Bad Charity

Fig. 2. Influence of construal level on number of matrices solved (control) or reportedto have been solved correctly in Study 2a as a function of construal level mindset(abstract or concrete) and payment condition (control, self-benefit only, self-benefitwith donation to good charity, self-benefit with donation to bad charity).


abstract (concrete) mindset were honest (dishonest) when onlythey themselves would gain from dishonesty.

In alignment with our earlier findings, when an abstract mind-set was activated, participants were dishonest to provide addi-tional support for the greater social good. When the charity wasgood, participants underreported the number of correctly solvedmatrices (M = 2.36) and hurt themselves financially which ensuredlarger donations relative to the control (M = 4.03), F(1,262) = 13.09,p = .001, g2

p = .173. Conversely, for the bad charity, participants inthe abstract mindset condition overreported the number of correctmatrices (M = 10.73) compared to the control (M = 4.03) andgained financially while simultaneously reducing donations to acharity that threatened the environment, F(1,262) = 40.96,p < .001, g2

p = .394.With the activation of a concrete mindset, self-gain occurred

across conditions in which dishonesty was possible. Participantsreported solving a significantly higher number of matrices andgained financially compared to the control condition (M = 4.06)when the charity was good (M = 6.95), F(1,262) = 18.58, p < .001,g2

p = .210, and when it was bad (M = 8.47), F(1,262) = 21.61,p < .001, g2

p = .255, suggesting that participants were dishonestfor personal gain. Together, these findings support our argumentthat activating an abstract mindset results in dishonesty in thedirection that does the most to support the greater good but whenactivating a concrete mindset, dishonesty consistently occurs inthe direction that leads to the largest personal financial gain.

Finally, we assessed the proportion of participants in each con-dition who reported solving (actually solved for control conditions)a low (0–2), average (3–5), high (6–10), or very high (11–20)

Table 2Influence of construal level on the proportion of participants who solved (control) or reporconstrual level mindset (abstract or concrete) and payment condition (control, self-beneficharity).

Condition

Number of solved matrices Concretecontrola

(n = 35)

Abstractcontrola

(n = 35)

Concreteself only(n = 33)

AbstracSelf onl(n = 34)

0–2 Matrices 20.0% 22.9% 9.1% 23.5%3–5 Matrices 62.9% 60.0% 36.4% 55.9%6–10 Matrices 17.1% 17.1% 42.4% 20.6%11–20 Matrices 0.0% 0.0% 12.1% 0.0%

Mean matrices solved 4.06 4.03 6.42 4.12

a In the control conditions, dishonesty was not possible.

number of matrices (see Table 2). We found that when an abstractmindset was activated, 58.4% of participants reported solving a lownumber of matrices when unearned money was donated to theMake-A-Wish Foundation while 80.0% reported solving six or morematrices when unearned money was said to be donated to the Re-source Use Fund. When a concrete mindset was activated, how-ever, the proportion of participants who reported solving six ormore matrices was relatively high whether the decision influencedthe amount donated to the Make-A-Wish Foundation (59.4%) orthe Resource Use Fund (66.7%). That is, consistent with the findingsfrom Study 1, many participants appear to have been slightly dis-honest when given the opportunity to do so, although the directionvaried based on whether an abstract or a concrete mindset wasactivated.

Consistent with and extending the results from Study 1, thefindings from Study 2a demonstrate that when primed with an ab-stract mindset, people generally abide by the principle of honestywhen only they would benefit from dishonesty but they are moreinclined to be dishonest when doing so further supports a greatersocial good. Specifically, we find that those who have an abstractmindset activated dishonestly underreport performance, whichenables them to lend additional support to sick children eventhough it simultaneously lowers their personal earnings. They willalso dishonestly overreport performance, which withholds dona-tions from a bad charity that threatens the environment whilesimultaneously increasing their own earnings. In other words,when faced with decisions involving principles that are intrinsi-cally linked such that violating one principle facilitates additionalsupport for another principle, one principle is prioritized over theother. In our study, we find support for the notion that when an ab-stract mindset is activated, people prioritize the desire to providesupport for the greater social good inferred from the situation overthe desire to remain honest and this occurs whether it simulta-neously results in higher or lower personal earnings.

When a concrete mindset was primed, however, the directionand degree of dishonesty did not vary substantially by condition.Instead, when dishonesty was possible, participants primed witha concrete mindset consistently overreported performance and in-creased personal earnings even when it simultaneously lowereddonations to an admired or unadmired charity. Together, thesefindings provide support for our proposition that a concrete mind-set will lead to a focus on subordinate, personal gain, while anabstract mindset will lead one to prioritize honesty when only one-self would benefit from dishonesty but to prioritize the superordi-nate, greater social good over honesty when dishonesty benefitsthis higher-level good.

In Study 2b, we further assess the different decision-makingprocesses between those who have an abstract or a concrete mind-set activated when the situation involves a potentially harmfulcharity. While there was a potential to use thought protocols

ted to have solved correctly a certain number of matrices in Study 2a as a function oft only, self-benefit with donation to good charity, self-benefit with donation to bad

ty

Concrete goodcharity(n = 37)

Abstract goodcharity(n = 36)

Concrete badcharity(n = 30)

Abstract badcharity(n = 30)

0.0% 58.4% 3.3% 0.0%40.6% 33.3% 30.0% 20.0%45.9% 8.3% 43.4% 33.3%13.5% 0.0% 23.3% 46.7%

6.95 2.36 8.47 10.73


within Study 2a to examine whether abstract and concrete mind-sets led to different reasons for why participants made a particulardecision, a concern was that the responses would have reflectedjustification of one’s actions and participants would not have beenwilling to admit to dishonesty for personal gain. To control for thispotential issue, we instead have participants judge the actions ofothers who made a similar decision within a hypothetical context.This allows us to examine how people who have an abstract orconcrete mindset activated can focus on different ends, whichinfluences how they make their decisions when confronted withpersonal or donation-related decisions.

Study 2b

Our findings suggest that with a concrete mindset, participantsprioritize subordinate, personal gain over honesty but that with anabstract mindset, participants prioritize the superordinate, greatersocial good over honesty when dishonesty can benefit oneself or agreater social good. That is, with a concrete mindset, overreportingoccurred whenever dishonesty was possible, suggesting that thepriority was on personal monetary gain. Similarly, with an abstractmindset, in Study 2a there was a tendency to underreport perfor-mance when the charity was good but to overreport when thecharity was bad. With the good charity, it is unlikely that underre-porting occurred due to a desire to hurt oneself financially but in-stead, it likely occurred out of a desire to support the greater goodinferred by the situation. However, when the charity was bad, par-ticipants in both the abstract and the concrete mindset conditionsoverreported performance.

With an abstract mindset, we posit that people focused on sup-porting the greater social good inferred from the situation andoverreported in the bad charity condition in order to withholdmoney from the charity that threatened the environment. Whereaswith a concrete mindset, we posit that the emphasis was on per-sonal gain, which was also accomplished by overreporting whenthe charity was bad. That is, while the outcome was the same(smaller donations but larger personal earnings), the emphasis thatled to those outcomes differed based on construal level. To assessthis more closely, we conducted Study 2b.

In the current study, we explore whether people primed withan abstract versus a concrete mindset differ in their perceptionsof why one would overreport when the charity is good or badbased on a scenario reflective of Study 2a. Our primary interestcenters on the situation involving the bad charity, but we includedthe good charity condition as a point of comparison. When thecharity is good, we anticipate that people primed with an abstractor concrete mindset will not differ in the extent to which they per-ceive that overreporting occurred out of a desire for personal gainrather than a desire to withhold money from the charity. However,when the charity is bad, we anticipate that by activating an ab-stract as opposed to a concrete mindset, people will suggest thatoverreporting occurred more from a desire to withhold moneyfrom the charity rather than out of a desire to increase personalearnings.

8 Participants also answered two additional questions based on the scale end-points used in the more parsimonious question reported in the main text. In onequestion, they indicated the extent to which they would say that the person reportedsolving more than they actually did because they wanted to get more money forthemselves and in the other question because they wanted to withhold or keep somemoney away from the Make-A-Wish Foundation (Resource Use Fund). Both questionswere asked on a scale from 1 (not at all) to 9 (completely).

9 Responses superordinate to the preceding ones were coded 1 and subordinateresponses were coded �1. Responses not fitting either criterion were coded 0.

Method


Two hundred two participants (mean age 34.99 years, 45.0% fe-male) took part in the study online and were randomly assigned toone of four between-participants conditions in a 2 (construal levelprime: abstract vs. concrete) � 2 (charity: good vs. bad) design. Asin the previous study, the good charity was the Make-A-WishFoundation and the bad charity was the fictitious Resource Use

Fund. After completing the construal level prime, participants reada scenario reflective of the previous study and indicated the extentto which they believed overreporting occurred from a desire to in-crease personal profits or to withhold money from the charity.

Procedure

Participants were first primed with either an abstract or a con-crete mindset with the same priming method (Freitas, Gollwitzer,& Trope, 2004) used in Study 2a. Next, participants read a scenariobased on the previous study. They imagined that a person did atask that involved answering 20 questions and that the personwould receive 25 cents for each correct answer. The person inthe scenario graded their own answers and destroyed the paperso nobody else would know how well they had actually done. Next,participants read that before the person received payment for thetask, they found out that the maximum payment of $5 was setaside and the person in the scenario would receive 25 cents foreach solved puzzle but the charity would receive the remainingportion of the $5 that the person did not receive. The charity waseither the Make-A-Wish Foundation (good) or the Resource UseFund (bad) and they read a brief description of the charity. Theparticipants then read that the person could report the numberof puzzles they actually solved (some money for themselves andthe charity), a higher number than they actually solved (moremoney for themselves and less for the charity), or a smaller num-ber than they actually solved (less money for themselves and morefor the charity). Next, participants read that the person in the sce-nario decided to report solving more puzzles than they actually didwhich would result in more money for themselves and a smallerdonation. Finally, participants indicated why the person made thisdecision on a nine-point scale where the low end was labeled ‘‘toget more money for themselves’’ and the high end was labeled‘‘to withhold or keep some money from the Make-A-Wish Founda-tion (Resource Use Fund).8’’ The participants in the bad charity con-dition then read that the Resource Use Fund is not real and allparticipants answered demographic questions.


Manipulation checks

As in the previous study, two independent judges coded the‘‘why’’ and ‘‘how’’ responses (Fujita, Trope, et al., 2006).9 Disagree-ments were resolved by discussion (initial agreement was 96.8%).The analysis revealed that the prime successfully manipulated con-strual level. Participants who responded to ‘‘why’’ prompts gener-ated statements that reflected an abstract mindset (M = 3.57)compared to those who responded to the ‘‘how’’ prompts(M = �3.66), t(200) = 75.27, p < .001.

Focus of overreporting decision

An ANOVA revealed a significant interaction F(1,198) = 6.91,p < .01, g2

p = .034, and significant main effects of both construal le-vel, F(1,198) = 10.40, p < .01, g2

p = .050, and charity conditionF(1,198) = 109.30, p < .001, g2

p = .356. We then assessed our


planned contrasts to gain further insight into the focus emphasizedby participants. As anticipated, when the charity was good (Make-A-Wish Foundation), the overreporting was viewed as being doneprimarily out of a desire to get more money for oneself and thisdid not significantly differ based on whether an abstract(M = 1.42) or a concrete (M = 1.23), F(1,198) = 1.10, p = .297, mind-set was activated. However, when the charity was bad (ResourceUse Fund), there was a significant difference between the viewsfor those primed with an abstract versus a concrete mindset.Specifically, we find that those with an abstract mindset perceivethat the person was more likely to have overreported due to a de-sire to withhold money from the bad charity (M = 5.50) than werethe participants primed with a concrete mindset (M = 3.67),F(1,198) = 9.53, p < .01, g2

p = .087.10

When the charity was bad, those with an abstract mindset weremore inclined to feel that overreporting occurred out of a desire towithhold money from the charity that threatened the greater goodinferred from the scenario whereas those with a concrete mindsetwere more likely to believe that overreporting was driven by a de-sire to increase personal gain. These findings in conjunction withthe behavioral findings across conditions within the prior studiessuggest that with an abstract mindset, people prioritize the super-ordinate, higher-level desire to support the greater social goodover honesty whereas with a concrete mindset, people prioritizethe subordinate, lower-level desire for personal gain over honesty.

General discussion and implications

Across these studies, our findings corroborate the proposal thatwhen only one ethical principle is at play, when an abstract mind-set is activated, people tend to abide by the ethical principle ofhonesty rather than on increasing personal gain. However, whentwo ethical principles conflict such that violating the principle ofhonesty will simultaneously do even more to help a greater socialgood, people must choose whether to remain honest or to be dis-honest in order to provide additional support for the greater socialgood. We find that with an abstract mindset, people prioritize thedesire to do even more for the greater social good over honesty. Wefeel that this provides a reasonable framework from which to con-sider the question of when the greater social good is seen as asuperordinate, focal concern that overrides other ethical principlessuch as honesty.

Theoretical and practical implications

Our results contribute by extending the research on construallevel theory in relation to ethical decision-making when more thanone ethical principle can be viewed as being in conflict. Previousresearch suggests that an abstract, high-level construal leads tostrict adherence to moral principles (Agerström & Björklund,2009; Eyal et al., 2008). Much of the past work, however, focuseson judgments of the morality of various scenarios rather than ac-tual decision-making. When past research tested actual decisions,the decision to violate or adhere to ethical principles only influ-enced oneself—there was no greater social good that could benefitmore from violating rather than adhering to ethical principles(Agerström & Björklund, 2009).

We find that when faced with a single ethical principle whenonly personal gain is at stake, people are inclined to adhere to thatethical principle when an abstract mindset is activated. However,our studies also extend the theory into previously overlooked deci-sion contexts and demonstrate that an abstract mindset does notalways lead people to abide by ethical principles, even when

10 Similar results were found when the two additional questions were analyzed.

adhering to the principles would still result in a positive outcome(albeit to a lesser extent). Instead, they behave dishonestly whendoing so provides more money to a charity that helps a greater so-cial good or less money to a charity that could potentially harm agreater social good. We find that in these contexts, where a desireto behave honestly conflicts with a desire to do more for a socialgood, that the greater social good takes priority whether the deci-sion simultaneously decreases, increases, or has no impact on one’spersonal earnings.

Oftentimes, conflicts between ethical principles are explored inthe context of true moral dilemmas. With moral dilemmas, anundesirable act must occur (e.g., death of one, stealing) in orderfor a desirable outcome to occur (e.g., saving many lives, healingone’s spouse) (Foot, 1967; Kohlberg, 1969). In our research, how-ever, we used situations where undesirable acts do not have to oc-cur for there to be desirable outcomes. Honesty would still result insome funds being donated (withheld) from an admirable (unad-mirable) charity but dishonesty had the potential to lead to evenmore funds being donated (withheld). Therefore, our research alsoadds to the literature by highlighting the value of situationalfactors that may strengthen the focus on a desire to support agreater social good and result in the violation of other ethical prin-ciples. We find that this occurs even when violating these ethicalprinciples is not a requirement of helping a greater good as is thecase in traditional moral dilemmas.

We also contribute to the literature in which people will over-look violating ethical principles such as honesty in order to achievedesirable outcomes. For instance, people who are dishonest arefound to be more forgetful of moral rules such as the items in anhonor code (Shu, Gino, & Bazerman, 2011) and are able to selec-tively overlook and self-justify the wrongness of their actions(Bandura, 1999; Bersoff, 1999). In research considering the rela-tionship between the ends and the means, it is found that whenanother person’s unethical behaviors result in positive outcomes,people judge those unethical acts less harshly (Gino & Pierce,2010; Gino, Shu, & Bazerman, 2010). Consistent with this past re-search, we show that construal level is another situational factorthat may influence whether people will focus more on the desir-able outcomes associated with a greater social good rather thanthe need to violate ethical principles in order to achieve thoseoutcomes.

Our findings also have practical implications for situationswhere unethical behaviors can support a greater social good andthe situational elements in the environment encourage construingthe event in more abstract forms. Decisions that influence a largersocial good may easily occur in the distant rather than the near fu-ture. Greater temporal distance is associated with an abstract,high-level construal while temporal closeness is associated witha concrete, low-level construal (Trope & Liberman, 2003). Thus,this distant temporal orientation may activate an abstract constru-al of the overall situation. An environmental activist, for instance,may focus on the long-term, eventual impact of pollution on theenvironment and this distant, abstract orientation may result inan emphasis on the superordinate, ultimate social good (savingthe environment) which overshadows the subordinate, proximatemeans (vandalism) undertaken to serve the greater social good.

On a more day-to-day basis, distant locations such as anotherpart of the country or world are construed at more abstract levelsthan near locations (Fujita, Henderson, et al., 2006). Thus, whendecisions highlight physical distance and the associated abstractconstructs, it could lead to increased unethical behaviors under-taken as a conduit for the greater social good inferred from the sit-uation. A disgruntled employee reviewing expense reports, forexample, may be more inclined to support the cause of protectingworkers’ rights by overlooking otherwise questionable expenseswhen the expense report comes from the firm’s distant rather than


near offices. As administrative functions become more centralized,the abstractness associated with distant branches reporting to thehead-office could have an increasing influence on such decisions.

References

Agerström, J., & Björklund, F. (2009). Moral concerns are greater for temporallydistant events and are moderated by value strength. Social Cognition, 27,261–282.

Amit, E., & Greene, J. D. (2012). You see, the ends don’t justify the means: Visualimagery and moral judgment. Psychological Science, 23, 861–868.

Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities.Personality and Social Psychology Review, 3, 193–209.

Bartels, D. M. (2008). Principled moral sentiment and the flexibility of moraljudgment and decision making. Cognition, 108, 381–417.

Batson, C., Batson, J. G., Todd, R., Brummett, B. H., Shaw, L. L., & Aldeguer, C. R.(1995). Empathy and the collective good: Caring for one of the others in a socialdilemma. Journal of Personality and Social Psychology, 68, 619–631.

Batson, C., Klein, T. R., Highberger, L., & Shaw, L. L. (1995). Immorality fromempathy-induced altruism: When compassion and justice conflict. Journal ofPersonality and Social Psychology, 68, 1042–1054.

Bersoff, D. M. (1999). Why good people sometimes do bad things: Motivatedreasoning and unethical behavior. Personality and Social Psychology Bulletin, 25,28–39.

Broeders, R., van den Bos, K., Müller, P. A., & Ham, J. (2011). Should I save or should Inot kill? How people solve moral dilemmas depends on which rule is mostaccessible. Journal of Experimental Social Psychology, 47, 923–934.

Ditto, P. H., & Liu, B. (2012). Deontologial dissonance and the consequentialistcrutch. In M. Mikulincer & P. R. Shaver (Eds.), The social psychology of morality:Exploring the causes of good and evil (pp. 51–70). Washington, DC: AmericanPsychological Association.

Donaldson, T., & Dunfee, T. W. (1994). Toward a unified conception of businessethics: Integrative social contracts theory. The Academy of Management Review,19, 252–284.

Erat, S., & Gneezy, U. (2012). White lies. Management Science, 58, 723–733.Eyal, T., & Liberman, N. (2012). Morality and psychological distance: A construal

level theory perspective. In M. Mikulincer & P. R. Shaver (Eds.), The socialpsychology of morality: Exploring the causes of good and evil (pp. 185–202).Washington, DC: American Psychological Association.

Eyal, T., Liberman, N., & Trope, Y. (2008). Judging near and distant virtue and vice.Journal of Experimental Social Psychology, 44, 1204–1209.

Foot, P. (1967). The problem of abortion and the doctrine of double effect. OxfordReview, 5, 5–15.

Förester, J., Friedman, R. S., & Liberman, N. (2004). Temporal construal effects onabstract and concrete thinking: Consequences for insight and creativecognition. Journal of Personality and Social Psychology, 87, 177–189.

Freitas, A. L., Gollwitzer, P., & Trope, Y. (2004). The influence of abstract andconcrete mindsets on anticipating and guiding others’ self-regulatory efforts.Journal of Experimental Social Psychology, 40, 739–752.

Fujita, K., Henderson, M. D., Eng, J., Trope, Y., & Liberman, N. (2006). Spatial distanceand mental construal of social events. Psychological Science, 17, 278–282.

Fujita, K., Trope, Y., Liberman, N., & Levin-Sagi, M. (2006). Construal levels and self-control. Journal of Personality and Social Psychology, 90, 351–367.

Gino, F., & Pierce, L. (2009). Dishonesty in the name of equity. Psychological Science,20, 1153–1160.

Gino, F., & Pierce, L. (2010). Lying to level the playing field: Why people maydishonestly help or hurt others to create equity. Journal of Business Ethics, 95,89–103.

Gino, F., Shu, L. L., & Bazerman, M. H. (2010). Nameless + harmless = blameless:When seemingly irrelevant factors influence judgment of (un)ethical behavior.Organizational Behavior and Human Decision Processes, 111, 93–111.

Gneezy, U. (2005). Deception: The role of consequences. The American EconomicReview, 95, 384–394.

Hunt, C. V., Kim, A., Borgida, E., & Chaiken, S. (2010). Revisiting the self-interestversus values debate: The role of temporal perspective. Journal of ExperimentalSocial Psychology, 46, 1155–1158.

Isikoff, M. (2014). After 43 years, activists admit theft at FBI office that exposeddomestic spying. NBC News. <http://investigations.nbcnews.com/_news/2014 /01/06/22205443-after-43-years-activists-admit-theft-at-fbi-office-that-exposed-domestic-spying>.

Kivetz, Y., & Tyler, T. R. (2007). Tomorrow I’ll be me: The effect of time perspectiveon the activation of idealistic versus pragmatic selves. Organizational Behaviorand Human Decision Processes, 102, 193–211.

Kohlberg, L. (1969). Stage and sequence: The cognitive-developmental approach tosocialization. In D. A. Goslin (Ed.), Handbook of socialization theory and research(pp. 347–489). Chicago, IL: Rand McNally.

Liberman, N., Sagristano, M. D., & Trope, Y. (2002). The effect of temporal distanceon level of mental construal. Journal of Experimental Social Psychology, 38,523–534.

Liberman, N., & Trope, Y. (1998). The role of feasibility and desirabilityconsiderations in near and distant future decisions: A test of temporalconstrual theory. Journal of Personality and Social Psychology, 75, 5–18.

Mazar, N., Amir, O., & Ariely, D. (2008). The dishonesty of honest people: A theory ofself-concept maintenance. Journal of Marketing Research, 45, 633–644.

Mead, N. L., Baumeister, R. F., Gino, F., Schweitzer, M. E., & Ariely, D. (2009). Tootired to tell the truth: Self-control resource depletion and dishonesty. Journal ofExperimental Social Psychology, 45, 594–597.

Sagristano, M. D., Trope, Y., & Liberman, N. (2002). Time-dependent gambling: Oddsnow, money later. Journal of Experimental Psychology: General, 131, 364–376.

Scott, E. D. (2003). Plane truth: A qualitative study of employee dishonesty in theairline industry. Journal of Business Ethics, 42, 321–337.

Shalvi, S., Dana, J., Handgraaf, M. J. J., & De Dreu, C. K. W. (2011). Justified ethicality:Observing desired counterfactuals modifies ethical perceptions and behavior.Organizational Behavior and Human Decision Processes, 115, 181–190.

Shu, L. L., Gino, F., & Bazerman, M. H. (2011). Dishonest deed, clear conscience:When cheating leads to moral disengagement and motivated forgetting.Personality and Social Psychology Bulletin, 37, 330–349.

Stephan, E., Liberman, N., & Trope, Y. (2010). Politeness and psychological distance:A construal level perspective. Journal of Personality and Social Psychology, 98,268–280.

Thomson, J. J. (1985). The trolley problem. Yale Law Journal, 94, 1395–1415.Toure-Tillery, M., & Fishbach, A. (2012). The end justifies the means, but only in the

middle. Journal of Experimental Psychology: General, 141, 570–583.Trope, Y., & Liberman, N. (2003). Temporal construal. Psychological Review, 110,

403–421.Umphress, E. E., & Bingham, J. B. (2011). When employees do bad things for good

reasons: Examining unethical pro-organizational behaviors. OrganizationScience, 22, 621–640.

Vardi, Y., & Wiener, Y. (1996). Misbehavior in organizations: A motivationalframework. Organizational Science, 7, 151–165.

Vinograd, C. (2011). ‘Anonymous’ hackers target US security think tank. AssociatedPress. <http://www.msnbc.msn.com/id/45787767/ns/technology_and_science-security/t/anonymous-hackers-target-us-security-think-tank>.

http://refhub.elsevier.com/S0749-5978(14)00010-7/h0005
































































http://investigations.nbcnews.com/_news/2014%20/01/06/22205443-after-43-years-activists-admit-theft-at-fbi-office-that-exposed-domestic-spying?lite











































http://www.msnbc.msn.com/id/45787767/ns/technology_and_science-security/t/anonymous-hackers-target-us-security-think-tank

http://www.msnbc.msn.com/id/45787767/ns/technology_and_science-security/t/anonymous-hackers-target-us-security-think-tank

Documents

Ethical ends: Effect of abstract mindsets in ethical decisions for the greater social good